I would run it without any charges added and measure the MullikenPopulations on the atoms. Unless there is a clear physical mechanisms that would be responsible for binding charges to a particular atoms (or you have an isolated system) it is not physical to impose a particular charge state, since the system is full of free mobile charges that can move, and will do so, to find the state of lowest total energy as the solution. This process might end up "charging" the cluster compared to it's isolated state, but it's not like a 3+ charged isolated ionic system just sitting on top of the 2D sheet - as long there is some chemical bonding between those two systems, any mobile charge will just move out of the cluster and into the infinite leads if that is energetically favorable. So the only reason to introduce opposing compensation charges to bind electrons to particular positions in space is if there is a real physical reason. This is why we use this for doping - the dopants are the physical reason, but we don't model the actual dopant atoms (As in Si, for instance) since it's hard to control the concentration (only possible to model multiples of 1/N, where N is the number of atoms). So instead we put a little amount of "dopant effect" on each Si atom (i.e. we smear out the dopant core charge over many atoms) to attract or repel electrons. But again, there is a real physical reason for this (the dopants). I don't see that in your case.
